Absence of Nuclear p16 Is a Diagnostic and Independent Prognostic Biomarker in Squamous Cell Carcinoma of the Cervix

Abstract

The tumor-suppressor protein p16 is paradoxically overexpressed in cervical cancer (CC). Despite its potential as a biomarker, its clinical value and the reasons for its failure in tumor suppression remain unclear. Our purpose was to determine p16 clinical and biological significance in CC. p16 expression pattern was examined by immunohistochemistry in 78 CC cases (high-grade squamous intraepithelial lesions (HSILs) and squamous cell carcinomas of the cervix -SCCCs). CC cell proliferation and invasion were monitored by real-time cell analysis and Transwell^® invasion assay, respectively. Cytoplasmic p16 interactors were identified from immunoprecipitated extracts by liquid chromatography-tandem mass spectrometry, and colocalization was confirmed by double-immunofluorescence. We observed that SCCCs showed significantly more cytoplasmic than nuclear p16 expression than HSILs. Importantly, nuclear p16 absence significantly predicted poor outcome in SCCC patients irrespective of other clinical parameters. Moreover, we demonstrated that cytoplasmic p16 interacted with CDK4 and other unreported proteins, such as BANF1, AKAP8 and AGTRAP, which could sequester p16 to avoid nuclear translocation, and then, impair its anti-tumor function. Our results suggest that the absence of nuclear p16 could be a diagnostic biomarker between HSIL and SCCC, and an independent prognostic biomarker in SCCC; and explain why p16 overexpression fails to stop CC growth.

Keywords: cervical cancer; cytoplasmic p16; high-grade squamous intraepithelial lesion; nuclear p16; predictive biomarker; squamous cell carcinoma of the cervix; subcellular location.

Figure 1

Global p16 expression in cervical tumors. (A) Immunohistochemical expression of p16 was examined in a series of 29 high-grade squamous intraepithelial lesions (HSILs) and 49 squamous cell carcinomas of the cervix (SCCCs) (*** p < 0.001). Representative images are shown at 200× magnification, and details are highlighted at 400×. (B) Association between global p16 expression and progression-free survival (PFS) and overall survival (OS) in SCCC patients. (C) Multivariate analysis showing the independent association between p16 expression and PFS (left) or OS (right), regardless of vascular invasion, stage and age of SCCC patients. (CI, confidence interval).

Figure 2

Subcellular location of p16 in cervical tumors. (A) Representative images of p16 staining in 3 high-grade intraepithelial lesions (HSILs) and 3 squamous carcinomas of the cervix (SCCCs) at 400× magnification showing that SCCC samples (n = 49) showed stronger cytoplasmic than nuclear p16 expression than HSIL (n = 29), regardless of the degree of global p16 staining (strong, moderate or low). Note p16 negativity in the nuclei of the low p16-expressing SCCC sample. The ratio cytoplasmic to nuclear (cyt/nuc) p16 expression was calculated by dividing p16 immunohistochemical score in each subcellular compartment. The horizontal line in the histogram shows equal amounts of cytoplasmic and nuclear p16 (ratio cyt/nuc p16 = 1) (*** p < 0.001). (B) Association between p16 subcellular location and overall survival (OS) in SCCC patients. (C) Multivariate analysis revealed an independent association between p16 subcellular location and OS, regardless of vascular invasion, stage and age of SCCC patients. (CI, confidence interval).

Figure 3

p16 subcellular location in SCCC cell lines. (A) Nuclear and cytoplasmic protein fractions of C-33A and SiHa were separately subjected to western blot to check p16 expression. GAPDH and Histone H3 were used as loading controls of each subcellular fraction. Numbers indicate the ratio of p16 signal relative to that in the loading control, measured by densitometry. (B) p16 expression was examined by immunofluorescence in C-33A and SiHa cells. Images were acquired at 400× magnification. (C) Cell proliferation of C-33A and SiHa cell lines was measured by real-time cell analysis for 7 days (* p < 0.05). (D) Cell invasion of C-33A and SiHa cell lines was examined by their ability to penetrate a Matrigel^® layer for 3 days. Images were acquired at 50× magnification.

Figure 4

Cytoplasmic p16 interactors in SiHa cells. (A) Native form of p16 protein was examined in both C-33A and SiHa cell lines by western blot under non-denaturing conditions. Arrows point to the proteins of interest, while the arrowhead indicates a different 3D conformation of native p16 in SiHa, as compared with C-33A. α-tubulin was used as a loading control. (B) Exclusively cytoplasmic p16 was immunoprecipitated from SiHa cells, and efficiency was checked by western blot. Immunoprecipitated (IP) and unbound (UB) fractions upon incubation with IgG and anti-p16 antibodies are shown. α-tubulin was used as a loading control. Immunoprecipitation of p16 interactors was checked by incubating the same membrane with an anti-CDK4 antibody, a very well-known p16-interacting protein. The arrows indicate the proteins of interest. (C) Colocalization in the cytoplasm of SiHa cells of p16 (in green) and four interactors (in red), identified by liquid chromatography-tandem mass spectrometry (LC-MS/MS), was revealed by double immunofluorescence. Images were acquired at 400× magnification.